| { |
| "format": "minijev-v2", |
| "config": { |
| "seed": 0, |
| "task": { |
| "name": "snake", |
| "choice_style": "semantic" |
| }, |
| "data": { |
| "dir": "runs/round2-dagger/data-it2" |
| }, |
| "env": { |
| "board_size": 10, |
| "reverse_action": "die", |
| "reward": { |
| "food": 1.0, |
| "death": -1.0, |
| "step": 0.0, |
| "win": 1.0 |
| } |
| }, |
| "model": { |
| "d_model": 96, |
| "n_heads": 4, |
| "n_layers": 3, |
| "n_decision_layers": 1, |
| "mlp_ratio": 2.0, |
| "dropout": 0.0, |
| "max_question_len": 32, |
| "max_choice_len": 16, |
| "state_encoders": { |
| "grid": { |
| "stem_channels": 32, |
| "patch_size": 1, |
| "max_grid": 34 |
| } |
| } |
| }, |
| "training": { |
| "batch_size": 256, |
| "epochs": 1, |
| "lr": 0.0005, |
| "min_lr_ratio": 0.05, |
| "weight_decay": 0.05, |
| "warmup_steps": 100, |
| "grad_clip": 1.0, |
| "choice_dropout": 0.1, |
| "eval_every": 1000, |
| "checkpoint_every": 1000, |
| "log_every": 100, |
| "eval_max_rows": 20000, |
| "hardware": { |
| "device": "auto", |
| "amp": "auto", |
| "num_workers": "auto", |
| "compile": "auto" |
| }, |
| "init_checkpoint": "runs/round2-dagger/it1/checkpoints/final.pt" |
| }, |
| "loss": { |
| "hard_weight": 1.0, |
| "label_smoothing": 0.0, |
| "soft_weight": 0.0, |
| "kd_weight": 1.0, |
| "kd_temperature": 2.0 |
| }, |
| "calibration": { |
| "method": "temperature" |
| }, |
| "eval": { |
| "games": 200, |
| "parallel": 64 |
| }, |
| "name": "snake-small-dagger8-dagger2", |
| "dagger": { |
| "keep_prob": 0.2, |
| "explore_eps": 0.02, |
| "epochs": 1, |
| "lr": 0.0005 |
| } |
| }, |
| "temperature": 1.4188514947891235, |
| "task_temperatures": {}, |
| "metrics": { |
| "avg_score": 88.56, |
| "win_rate": 0.735, |
| "games": 200 |
| } |
| } |
